Disaster-Responsive Fetal Movement Monitoring System for Flood Affected Rural Areas

Abstract

Climate-related weather disasters not only cause billions of dollars in damages but also have long-lasting effects on human health. In Year 2022, approximately 600,000 pregnant women were affected by devastating floods in Pakistan. The majority of these pregnant women belonged to rural areas and had limited access to healthcare facilities. Fetal movement serves as a reliable indicator of a healthy fetus. The proposed approach involves using an accelerometer measurement of fetal movement, along with a mobile application that allows for easy usage outside clinical environments, enabling remote fetal movement monitoring. By preprocessing a pre-recorded dataset of the 3D accelerometer measurements, four state-of-the-art machine learning algorithms are implemented to classify fetal movement with a relative degree of accuracy. The Extreme Gradient Boost algorithm demonstrates superior performance in classifying fetal movement, achieving an accuracy of 94.58% and an average accuracy of 87.03% through k-fold cross-validation.